LLMling Agent CLI¶

The LLMling Agent CLI provides a comprehensive set of commands to manage and interact with AI agents. It's designed around the concept of an "active agent file" - a YAML configuration that defines your agents and their settings. This avoids the need to pass the config file path each time you want to run a command.

Active Agent File¶

The CLI maintains an "active agent file" setting which determines which agents are available for commands like run, task, or watch. You can:

Add agent files with llmling-agent add <name> <path>
Set the active file with llmling-agent set <name>
List agents from the active config with llmling-agent list

Most commands will use the currently active agent file by default, but can be overridden with the --config option.

Available Commands¶

Agent Management¶

Command	Description
`add`	Register a new agent configuration file
`set`	Set the active configuration file
`list`	Show agents from the active (or specified) configuration

Execution¶

Command	Description
`run`	Run a node (agent/team) with prompts
`task`	Execute a defined task with an agent
`watch`	Run agents in event-watching mode

Server Commands¶

Command	Description
`serve-mcp`	Run agents as an MCP (Model Context Protocol) server
`serve-acp`	Run agents as an ACP (Agent Client Protocol) server
`serve-api`	Run agents as an OpenAI-compatible completions API server

History Management¶

Command	Description
`history show`	Show conversation history with filtering options
`history stats`	Show usage statistics
`history reset`	Reset (clear) conversation history

Quick Start¶

Add and activate an agent configuration:

llmling-agent add myconfig agents.yml
llmling-agent set myconfig

List available agents:
```
llmling-agent list
```

Run a prompt with an agent:

llmling-agent run analyzer "Analyze this text"

Command Examples¶

Running Agents¶

# Run a single agent with a prompt
llmling-agent run myagent "Analyze this"

# Run a team
llmling-agent run myteam "Process this"

# Show detailed output with costs
llmling-agent run myagent "Hello" --detail full --costs

Executing Tasks¶

# Execute a defined task
llmling-agent task docs write_api_docs

# Execute with additional prompt
llmling-agent task docs write_api_docs --prompt "Include code examples"

Server Commands¶

# Run as MCP server (stdio transport)
llmling-agent serve-mcp config.yml

# Run as MCP server with SSE transport
llmling-agent serve-mcp config.yml --transport sse --port 3001

# Run as ACP server for desktop integration
llmling-agent serve-acp config.yml

# Run as OpenAI-compatible API server
llmling-agent serve-api config.yml --port 8000

History Commands¶

# Show last 5 conversations
llmling-agent history show -n 5

# Show conversations from last 24 hours
llmling-agent history show --period 24h

# Show stats grouped by model
llmling-agent history stats --period 1w --group-by model

# Clear history for specific agent
llmling-agent history reset --agent myagent

Global Options¶

Option	Description
`--log-level`, `-l`	Set log level (default: info)
`--help`	Show help message

Configuration Files¶

Agent configurations are YAML files that define:

Available agents and their capabilities
System prompts and knowledge sources
Tool configurations
Response types
And more

Example:

agents:
  analyzer:
    name: "Text Analyzer"
    model: openai:gpt-4o
    description: "Analyzes text and provides structured output"
    toolsets:
      - type: file_access

See the Configuration Guide for detailed information about agent configuration.